Embed Size (px)
Citation preview
Sadie Larsen PhD Clement J Zablocki VA
Medical College of Wisconsin
Shannon Wiltsey-Stirman PhD National Center for PTSD
Stanford University
Maggi Mackintosh PhD National Center for PTSD
Wyatt Evans Michael E DeBakey VA Medical Center
Candice Monson PhD Ryerson University
Patricia Resick PhD Duke University
Poll Question 1 What is your primary role in the VA
Student trainee or fellow
Clinician
Researcher
Administrator manager or policy-maker
Other
Trauma-focused treatments Front-line treatment for PTSD
Under-utilized relative to their efficacy and the prevalence of PTSD
The concern with exacerbations Some populations (eg CSA comorbidities) canrsquot tolerate trauma-focused treatment
Trauma-focused treatments will make PTSD worse
Trauma-focused treatments could increase patient distress
Patients will then drop out or be worse off than when they started
Past Studies Imaginal exposure not linked to exacerbations (Foa et al
2002)
Drop-out rates same across active PTSD treatments (Hembree et al 2003)
Two studies of pre-to-post treatment worsening Some got worse on the wait list none got worse in trauma-
focused treatment (Jayawickreme et al 2013 Ehlers et al 2014)
Two studies of within-treatment exacerbations
30 of sample in an active trauma group (Mott et al 2013)
22 of sample had depression spikes (Keller et al 2014)
An examination of symptom exacerbations in a clinical trial sample (Larsen et al 2015)
Study 1 Questions 1 How common are symptom exacerbations in
trauma-focused treatments for PTSD
2 What predicts symptom exacerbations
3 Do symptom exacerbations predict worse post-treatment outcomes or dropout
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments
Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
MethodsmdashParticipants Two RCTs of CBT for PTSD
CPT
CPT-C
PE
Female survivors of interpersonal violence
Completed at least 4 therapy sessions
Demographics N = 192 (PE = 60 CPT = 98 CPT-C = 34)
Age M = 34 years
78 White 19 African-American
Marital status
44 single
25 married or cohabiting
30 separated widowed or divorced
Years since assault M = 11
Methods - Treatments Prolonged Exposure (PE)ndash 9 sessions
Psychoeducation
Breathing retraining
In vivo exposure
Imaginal exposure and emotional processing
Cognitive Processing Therapy (CPT) ndash 12 sessions Recognizing and challenging dysfunctional trauma-
related beliefs
Write trauma narrative
CPT-C does not include the written narrative
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Poll Question 1 What is your primary role in the VA
Student trainee or fellow
Clinician
Researcher
Administrator manager or policy-maker
Other
Trauma-focused treatments Front-line treatment for PTSD
Under-utilized relative to their efficacy and the prevalence of PTSD
The concern with exacerbations Some populations (eg CSA comorbidities) canrsquot tolerate trauma-focused treatment
Trauma-focused treatments will make PTSD worse
Trauma-focused treatments could increase patient distress
Patients will then drop out or be worse off than when they started
Past Studies Imaginal exposure not linked to exacerbations (Foa et al
2002)
Drop-out rates same across active PTSD treatments (Hembree et al 2003)
Two studies of pre-to-post treatment worsening Some got worse on the wait list none got worse in trauma-
focused treatment (Jayawickreme et al 2013 Ehlers et al 2014)
Two studies of within-treatment exacerbations
30 of sample in an active trauma group (Mott et al 2013)
22 of sample had depression spikes (Keller et al 2014)
An examination of symptom exacerbations in a clinical trial sample (Larsen et al 2015)
Study 1 Questions 1 How common are symptom exacerbations in
trauma-focused treatments for PTSD
2 What predicts symptom exacerbations
3 Do symptom exacerbations predict worse post-treatment outcomes or dropout
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments
Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
MethodsmdashParticipants Two RCTs of CBT for PTSD
CPT
CPT-C
PE
Female survivors of interpersonal violence
Completed at least 4 therapy sessions
Demographics N = 192 (PE = 60 CPT = 98 CPT-C = 34)
Age M = 34 years
78 White 19 African-American
Marital status
44 single
25 married or cohabiting
30 separated widowed or divorced
Years since assault M = 11
Methods - Treatments Prolonged Exposure (PE)ndash 9 sessions
Psychoeducation
Breathing retraining
In vivo exposure
Imaginal exposure and emotional processing
Cognitive Processing Therapy (CPT) ndash 12 sessions Recognizing and challenging dysfunctional trauma-
related beliefs
Write trauma narrative
CPT-C does not include the written narrative
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Trauma-focused treatments Front-line treatment for PTSD
Under-utilized relative to their efficacy and the prevalence of PTSD
The concern with exacerbations Some populations (eg CSA comorbidities) canrsquot tolerate trauma-focused treatment
Trauma-focused treatments will make PTSD worse
Trauma-focused treatments could increase patient distress
Patients will then drop out or be worse off than when they started
Past Studies Imaginal exposure not linked to exacerbations (Foa et al
2002)
Drop-out rates same across active PTSD treatments (Hembree et al 2003)
Two studies of pre-to-post treatment worsening Some got worse on the wait list none got worse in trauma-
focused treatment (Jayawickreme et al 2013 Ehlers et al 2014)
Two studies of within-treatment exacerbations
30 of sample in an active trauma group (Mott et al 2013)
22 of sample had depression spikes (Keller et al 2014)
An examination of symptom exacerbations in a clinical trial sample (Larsen et al 2015)
Study 1 Questions 1 How common are symptom exacerbations in
trauma-focused treatments for PTSD
2 What predicts symptom exacerbations
3 Do symptom exacerbations predict worse post-treatment outcomes or dropout
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments
Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
MethodsmdashParticipants Two RCTs of CBT for PTSD
CPT
CPT-C
PE
Female survivors of interpersonal violence
Completed at least 4 therapy sessions
Demographics N = 192 (PE = 60 CPT = 98 CPT-C = 34)
Age M = 34 years
78 White 19 African-American
Marital status
44 single
25 married or cohabiting
30 separated widowed or divorced
Years since assault M = 11
Methods - Treatments Prolonged Exposure (PE)ndash 9 sessions
Psychoeducation
Breathing retraining
In vivo exposure
Imaginal exposure and emotional processing
Cognitive Processing Therapy (CPT) ndash 12 sessions Recognizing and challenging dysfunctional trauma-
related beliefs
Write trauma narrative
CPT-C does not include the written narrative
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
The concern with exacerbations Some populations (eg CSA comorbidities) canrsquot tolerate trauma-focused treatment
Trauma-focused treatments will make PTSD worse
Trauma-focused treatments could increase patient distress
Patients will then drop out or be worse off than when they started
Past Studies Imaginal exposure not linked to exacerbations (Foa et al
2002)
Drop-out rates same across active PTSD treatments (Hembree et al 2003)
Two studies of pre-to-post treatment worsening Some got worse on the wait list none got worse in trauma-
focused treatment (Jayawickreme et al 2013 Ehlers et al 2014)
Two studies of within-treatment exacerbations
30 of sample in an active trauma group (Mott et al 2013)
22 of sample had depression spikes (Keller et al 2014)
An examination of symptom exacerbations in a clinical trial sample (Larsen et al 2015)
Study 1 Questions 1 How common are symptom exacerbations in
trauma-focused treatments for PTSD
2 What predicts symptom exacerbations
3 Do symptom exacerbations predict worse post-treatment outcomes or dropout
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments
Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
MethodsmdashParticipants Two RCTs of CBT for PTSD
CPT
CPT-C
PE
Female survivors of interpersonal violence
Completed at least 4 therapy sessions
Demographics N = 192 (PE = 60 CPT = 98 CPT-C = 34)
Age M = 34 years
78 White 19 African-American
Marital status
44 single
25 married or cohabiting
30 separated widowed or divorced
Years since assault M = 11
Methods - Treatments Prolonged Exposure (PE)ndash 9 sessions
Psychoeducation
Breathing retraining
In vivo exposure
Imaginal exposure and emotional processing
Cognitive Processing Therapy (CPT) ndash 12 sessions Recognizing and challenging dysfunctional trauma-
related beliefs
Write trauma narrative
CPT-C does not include the written narrative
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Past Studies Imaginal exposure not linked to exacerbations (Foa et al
2002)
Drop-out rates same across active PTSD treatments (Hembree et al 2003)
Two studies of pre-to-post treatment worsening Some got worse on the wait list none got worse in trauma-
focused treatment (Jayawickreme et al 2013 Ehlers et al 2014)
Two studies of within-treatment exacerbations
30 of sample in an active trauma group (Mott et al 2013)
22 of sample had depression spikes (Keller et al 2014)
An examination of symptom exacerbations in a clinical trial sample (Larsen et al 2015)
Study 1 Questions 1 How common are symptom exacerbations in
trauma-focused treatments for PTSD
2 What predicts symptom exacerbations
3 Do symptom exacerbations predict worse post-treatment outcomes or dropout
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments
Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
MethodsmdashParticipants Two RCTs of CBT for PTSD
CPT
CPT-C
PE
Female survivors of interpersonal violence
Completed at least 4 therapy sessions
Demographics N = 192 (PE = 60 CPT = 98 CPT-C = 34)
Age M = 34 years
78 White 19 African-American
Marital status
44 single
25 married or cohabiting
30 separated widowed or divorced
Years since assault M = 11
Methods - Treatments Prolonged Exposure (PE)ndash 9 sessions
Psychoeducation
Breathing retraining
In vivo exposure
Imaginal exposure and emotional processing
Cognitive Processing Therapy (CPT) ndash 12 sessions Recognizing and challenging dysfunctional trauma-
related beliefs
Write trauma narrative
CPT-C does not include the written narrative
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
An examination of symptom exacerbations in a clinical trial sample (Larsen et al 2015)
Study 1 Questions 1 How common are symptom exacerbations in
trauma-focused treatments for PTSD
2 What predicts symptom exacerbations
3 Do symptom exacerbations predict worse post-treatment outcomes or dropout
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments
Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
MethodsmdashParticipants Two RCTs of CBT for PTSD
CPT
CPT-C
PE
Female survivors of interpersonal violence
Completed at least 4 therapy sessions
Demographics N = 192 (PE = 60 CPT = 98 CPT-C = 34)
Age M = 34 years
78 White 19 African-American
Marital status
44 single
25 married or cohabiting
30 separated widowed or divorced
Years since assault M = 11
Methods - Treatments Prolonged Exposure (PE)ndash 9 sessions
Psychoeducation
Breathing retraining
In vivo exposure
Imaginal exposure and emotional processing
Cognitive Processing Therapy (CPT) ndash 12 sessions Recognizing and challenging dysfunctional trauma-
related beliefs
Write trauma narrative
CPT-C does not include the written narrative
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Study 1 Questions 1 How common are symptom exacerbations in
trauma-focused treatments for PTSD
2 What predicts symptom exacerbations
3 Do symptom exacerbations predict worse post-treatment outcomes or dropout
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments
Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
MethodsmdashParticipants Two RCTs of CBT for PTSD
CPT
CPT-C
PE
Female survivors of interpersonal violence
Completed at least 4 therapy sessions
Demographics N = 192 (PE = 60 CPT = 98 CPT-C = 34)
Age M = 34 years
78 White 19 African-American
Marital status
44 single
25 married or cohabiting
30 separated widowed or divorced
Years since assault M = 11
Methods - Treatments Prolonged Exposure (PE)ndash 9 sessions
Psychoeducation
Breathing retraining
In vivo exposure
Imaginal exposure and emotional processing
Cognitive Processing Therapy (CPT) ndash 12 sessions Recognizing and challenging dysfunctional trauma-
related beliefs
Write trauma narrative
CPT-C does not include the written narrative
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
MethodsmdashParticipants Two RCTs of CBT for PTSD
CPT
CPT-C
PE
Female survivors of interpersonal violence
Completed at least 4 therapy sessions
Demographics N = 192 (PE = 60 CPT = 98 CPT-C = 34)
Age M = 34 years
78 White 19 African-American
Marital status
44 single
25 married or cohabiting
30 separated widowed or divorced
Years since assault M = 11
Methods - Treatments Prolonged Exposure (PE)ndash 9 sessions
Psychoeducation
Breathing retraining
In vivo exposure
Imaginal exposure and emotional processing
Cognitive Processing Therapy (CPT) ndash 12 sessions Recognizing and challenging dysfunctional trauma-
related beliefs
Write trauma narrative
CPT-C does not include the written narrative
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Demographics N = 192 (PE = 60 CPT = 98 CPT-C = 34)
Age M = 34 years
78 White 19 African-American
Marital status
44 single
25 married or cohabiting
30 separated widowed or divorced
Years since assault M = 11
Methods - Treatments Prolonged Exposure (PE)ndash 9 sessions
Psychoeducation
Breathing retraining
In vivo exposure
Imaginal exposure and emotional processing
Cognitive Processing Therapy (CPT) ndash 12 sessions Recognizing and challenging dysfunctional trauma-
related beliefs
Write trauma narrative
CPT-C does not include the written narrative
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Methods - Treatments Prolonged Exposure (PE)ndash 9 sessions
Psychoeducation
Breathing retraining
In vivo exposure
Imaginal exposure and emotional processing
Cognitive Processing Therapy (CPT) ndash 12 sessions Recognizing and challenging dysfunctional trauma-
related beliefs
Write trauma narrative
CPT-C does not include the written narrative
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
MethodsmdashMeasures CAPS Pre- and post-treatment
PTSD Symptom ScalePosttraumatic Diagnostic Scale Pre post and weekly during treatment (every other
session)
Defining exacerbations
Change greater than 615 points on PDSPSS (Foa et al 2002)
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Results - Frequencies Frequency overall
CPT 286
CPT-C 147
PE 20
Frequency between sessions 2 and 4 CPT 134
CPT-C 29
PE 150
CPT vs CPT-C χ2(N = 131) = 289 p = 089
PE vs CPT-C χ2(N = 94) = 332 p = 068
CPT vs PE χ2(N = 157) = 008 p = 78
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Predictors of exacerbations Potential predictors
Demographics
Trauma-related variables
Treatment type
Diagnostic variables
Avoidance symptom cluster
None were significant predictors
Marginal significance
Childhood Sexual Abuse
Alcohol abuse
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Post-Treatment Outcomes Do exacerbations cause worse post-treatment
outcomes Yes and nohellip
Yes Those who experienced an exacerbation were more likely to retain a PTSD diagnosis and were likely to continue to have higher PTSD symptom scores over the course of treatment
No Those who experienced an exacerbation showed (large) significant pre-to-post treatment improvement ending with scores within non-PTSD population norms
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Dropout Unrelated to symptom exacerbations
Unrelated to PDSPSS early sessions
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Large exacerbations (2x) N=14 (7 of sample)
Slightly more likely to drop out
Comparable pre-to-post changes
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Conclusions from Study I A minority of patients experience symptom exacerbations
Exacerbations do not preclude positive outcomes
Clients can tolerate such treatments
Symptom exacerbations may be a normal part of treatment and are less common than sudden gains
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Poll Question 2 To what extent do these findings mirror your own
clinical experience Irsquove noticed symptom exacerbations like this AND worry
about them
Irsquove noticed symptom exacerbations like this and DONrsquoT worry about them
I havenrsquot noticed exacerbations like this
I avoid doing trauma-focused therapy because of worries about these exacerbations
Othernot applicable
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
CPT provided by newly-trained clinicians
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Study TeamInvestigators
Candice Monson PhD
Shannon Wiltsey Stirman PhD
Norman Shields PhD
Collaborators
Michael Suvak
Randi McCabe
David Ross
Project Directors
Jeanine Lane MA
Josh Deloriea BA Philippe Shnaider MA
Other Contributors
Meredith Landy MA Lindsey Torbit MA Stephanie Taillefer MA Shankari Sharma MA Iris Sijercic BA Tiffany Jenzer BA Marta Maslej BA Adrianna Tassone Esztella Vezer BA Natalie Whitney Bojana Kabranova Anne Wagner PhD C Psych Nicole Pukay-Martin PhD C Psych Tasoula Masina Steven Dufour Winnie Chen Kymm Bontrager Jessica Petritis Katie Siverts Patricia Resick PhD
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Design Hybrid-III design (Curran et al 2012) RCT of 3 different post-workshop consultation
strategies Fidelity assessment only (No consultation) Written feedback on a randomly selected session 6 months after workshop Potential to become a CPT Provider
Standard Consultation Weekly group consultation with a CPT expert Discussion of cases No use of work samples
Technology-enhanced Consultation Weekly group consultation with a CPT expert Review of segments of audio recorded sessions Review of worksheets stuck point logs etc
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Therapists and Settings N=134
Caseload with PTSD 23 Clinics
10 Operational Stress Injury Clinics 3 Canadian Forces Clinics
3 Hospitals (multiple sub-clinics at three of the clinics)
52
21
10
17 7 Community-based clinics
37 Private Practitioners (provide
services to Veterans)
78 Urban Clinics 16 Suburban 5 Rural
Mean Caseload 28 (SD=23)
lt25
26-50
51-75
76-100
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Years of Experience
Therapists Mean Age= 47 (SD=11)
72 Female 27 Male
Degree
41
35
10
4
52 had prior CBT supervision
PhDPsyD
Masterrsquos
Bachelorrsquos level
MD
11
21
33
33 lt5
6-10
10-20
gt20
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Client participants N=188 Age M=35 SD=11 Education M=12 SD=2 53 Female RaceEthnicity
88 White 4 Native Canadian 3 Asian 2 Black 3 Other lt1 HispanicLatino
Marital Status 36 SingleDivorcedWidowed 58 MarriedIn a Committed Relationship
Veteran Status 70 of males and 17 of females were in military or were veterans
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Client Diagnoses Diagnosis
PTSD 98
Major Depressive Disorder 56
Substance Abuse or Dependence 12
Anxiety Disorder 17
Bipolar Disorder 5
Eating Disorder 4
ADHD 5
Borderline PD 14
Other PD 10
No differences in client demographic or diagnostic characteristics between
conditions
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Study Procedures Therapists completed CPT workshop
Therapists in all condition knew fidelity would be rated
Those in consultation received weekly consultation for 6 months
Therapists delivered CPT
Enrolled clients completed PCL-IV at every session
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Symptom Exacerbation Defined as an increase greater than 571 on the PCL
between adjacent sessions
656 reported at least one instance of symptom exacerbations during treatment
Of those who reported them the average was 1 session with symptom exacerbations (range = 0 ndash 4)
Symptom exacerbations occurred on average in 14 of sessions in which PCLs were available (or median = 11 of available sessions)
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Exacerbations by Session 50
40
30
Males
Females20
10
0 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Why the difference Different measure 571 points for exacerbation
Differences in clinical training
Differences in discipline and training background
Very limited exclusion criteria
Differences in supervisionconsultation
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Fidelity
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Predictors of Exacerbation Predictors Model 1 Model 2 Model 3 Model 4
OR (95 CI) OR (95 CI) OR (95 CI) OR (95 CI)
Age 103 (99 106) 103 (99 106) 103 (99 106) 103 (99 106)
Gender 139 (61 319) 166 (66 418) 164 (65 417) 149 (56 393)
Education 102 (88 119) 103 (88 121) 103 (88 121) 103 (88 121)
Marital Status 120 (59 246) 94 (41 217) 94 (40 217) 92 (39 218)
Military Status 114 (49 265) 138 (54 350) 141 (55 361) 138 (54 354)
Depression na 139 (62 308) 138 (62 309) 137 (61 309)
Anx Disorder na 62 (21 179) 62 (21 182) 66 (21 205)
SU Disorder na 27 (08 90) 26 (07 89) 23 (06 83)
Personality DO na 61 (23 160) 60 (22 161) 59 (21 160)
Session 1 PCL na na 101 (97 104) 101 (97 105)
Consultation
None vs Tech na na na 49 (19 129)
None vs Standard na na na 45 (02 1390)
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
sc
ore
s)
tom
s (P
CL
S
ymp
SD
PT
Overall PTSD Change 65
60
55
50
45
40
35
30 Pre S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12
Assessment Occasion
No Tech Standard Overall
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Effect Sizes
CPT in RCTs No Tech- Standard Full Consultation enhanced Sample
1691 95 109 178 129 (95 CI=127-
211)
1Watts et al 2013
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Exacerbations predicting dropout
Participants with symptom exacerbations were significantly LESS likely to drop out of treatment prior to completing at least 8 sessions
Looking at sessions 1 ndash 7 individually the likelihood of treatment drop out was not related to the presence of symptom exacerbations all χ2 (1)lt 17 all ps gt 05
People who had an exacerbation in any given session were no more likely to drop out than they were to finish treatment
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Predictors of Symptom Change Predictors B (95 CI) Predictors B (95 CI)
Age 59 (-22 46) Age 67 (-20 40)
Gender -97 (-999 340) Gender -95 (-978 1341)
Education -43 (-141 90) Education -46 (-147 92)
Marital Status -88 (-808 310) Marital Status -103 (-846 266)
Military Status 31 (-536 737) Military Status 36 (-537 77 9)
Depression 66 (-391 780) Depression 75 (-381 841)
Anx Disorder -151 (-1393 189) Anx Disorder -159 (-1420 157)
SU Disorder -09 (-945 866) SU Disorder -17 (-961 810)
Personality DO -57 (-774 428) Personality DO -65 (-796 402)
Session 1 PCL -81 (-39 17) Session 1 PCL -89 (-40 15)
Consultation Condition Consultation Condition None vs Tech 28 (-576 769) None vs Tech 39 (-544 808)
None vs Standard -115 (-1071 285) None vs Standard -104 (-1033 319)
Early Exacerbations -24 (-441 346) Total Exacerbations 81 (-168 399)
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Exacerbations and Treatment Response ITT sample Mean PCL reduction of 1528 points
(d=129)
Mean post-treatment PCL-IV
No Exacerbations=390 (184)
Exacerbations=4522 (1569)
Number and presence of exacerbations did not predict treatment response (PCLlt50 at post-treatment)
χ2 (4)lt 397 p=41
66 of people who experienced an exacerbation had a PCL below 50 at session 12
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Worsening of Symptoms 54 reported worse PCLs scores at session 12 compared to session 1
(worse gt 571 symptom increase)
Number of symptom exacerbations in early sessions (sessions 1 ndash 5) did not predict overall worsening OR = 122 (40 368)
Effects on Symptom Trajectories
Using piece-wise latent growth curve model (sessions 1 ndash 5 vs 5 ndash 12)
Number of exacerbations in early sessions did not predict linear slope for PCL scores during session 6 ndash 12 standardized effect = -27 p = 19
Standardized effect estimates for symptom change during sessions 1-5 = -47 and during sessions 5 ndash 12 = -67 (both sig)
Consultation Condition did not predict significant differences for symptom change nor for number of exacerbations for sessions 1 ndash 5
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Conclusions
Exacerbations may be common in practice
They donrsquot mean people wonrsquot improve
We still know little about what predicts them
May be related to decreased avoidance or to non-treatment
related factors
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Clinical Considerations Important to differentiate between therapies that
produce symptom increases in the short term and
those that are truly harmful
Potential drawbacks of not engaging in trauma-
focused treatments
We can normalize symptom increases but reassure that clients that people still get better
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
Questions Comments Sadie E Larsen PhD
Sadielarsenvagov
Shannon Wiltsey Stirman PhD
Shannonwiltsey-stirmanvagov
Reference
Larsen Wiltsey Stirman Smith amp Resick (2016) Symptom exacerbations in trauma-focused treatments Associations with treatment outcome and non-completion Behavior Research and Therapy 77 68-77
